Project Summary

The UK rail system suffered from congestion and overcrowding, affecting 55% of passengers in 2019, with cancellations and delays costing the economy nearly £8M weekly, particularly in Northern England. This project aims to develop a novel machine vision system to identify congestion on trains accurately. The objective is to provide passengers with live and precise information about how crowded trains are, empowering them to plan their journeys ahead and change them. Additionally, disabled passengers and those with mobility aids can benefit by knowing available spaces and empty seats on board. The interdisciplinary team includes experts in Machine Learning and the UK railway industry from both academia (University of Brighton) and industry (Distributed Analytics).

Project Achievements

Distributed Analytics have delivered the first ever combined Object Detection and Human Recognition AI algorithm suitable for use at scale in rail passenger experience management. This lightweight solution is robust and reliable, with a performance accuracy of over 90%. A bespoke dataset was collated and labelled to train this Machine Vision AI, offering another first of its kind resource for AI innovators in the rail and transport space. This asset accelerates the development process for future passenger-level rail Computer Vision technologies – helping Distributed Analytics and other businesses to apply brand new AI techniques without significant spin up time.

Conclusions

Find My Way has fuelled two first of their kind innovations in the rail passenger experience AI space: a combined Object Detection and Human Recognition algorithm suitable for use at scale, and a bespoke labelled passenger and seating imagery dataset to accelerate future Machine Vision innovation. This has demonstrated the validity and value of applying Image Recognition AI to the rail congestion challenge, the necessary first step in an industry wide AI revolution process. Combined with intensive research into cutting edge AI practices, this work creates a knowledge and toolset resource able to support future innovation – illuminating the way forward for enhancing user experience across UK passenger rail.

Next Steps

The next steps are live environment testing and dataset enhancement with real UK rail CCTV footage. Subsequent development of the mobile application frontend will then take Find My Way beyond TRL 4 into a fully commercial product. DA have greatly benefitted from the networking and business guidance opportunities stemming from TRIG, helping the team to fo rge new connections and elevate business planning and strategy. Beyond providing the power to help innovative new products and technologies commercialize, TRIG brings significant value to any organization working to grow or connect with their peers across both the transport industry and academia.